Using the North Atlantic regional setup for other regions

[devyani2509]

How can we make changes in the NA regional model so that it can be used for other regions, such as, Indian Ocean ?
The forcing files used for the NA setup is it taken from any global dataset or model output ?

[dionhaefner]

Unfortunately there is no general recipe to prepare the initial condition / forcing files for other regions. It is the user's responsibility to create files with the appropriate format.

After you've done that, integration should be straightforward: Just swap out the external files and you should be good to go.

[Sougata18]

I see around 60 variables present in the forcing_nc4.nc file. Does the forcing file for other regions has to have this specific format ?
It would be very helpful if you could give some insight on this.

Also what about wind_energy.npy and tidal_energy.npy? There's no time dimension,so these are for initial condition only right ?

data1 = np.load('tidal_energy.npy')
data2 = np.load('wind_energy.npy')
print(data1.shape, data2.shape)

(87, 89) (87, 89)

[dionhaefner]

The setup file is responsible for parsing the forcings and initial conditions, so you can just look at north_atlantic.py to find out which variables are read when (look at uses of _read_data). And you're free to change the format (and reading logic) if you prefer to roll your own.

Now that I look at it, tidal_energy.npy and wind_energy.npy seem to be unused, so you can ignore those - will add a PR that removes those from the repo.

[Sougata18]

This is helpful, Thanks!

Could you please give some insight on sst_rest and sss_rest. Like how these have been calculated to incorporate in forcing_nc4.nc ?

[dionhaefner]

These variables (together with sst_clim and sss_clim) encode the relationship between heat / salinity forcing and air-sea gradients in the form of a restoring time scale. See e.g. here: https://journals.ametsoc.org/view/journals/phoc/28/5/1520-0485_1998_028_0890_ohtstb_2.0.co_2.xml

For reference, total heat forcing is computed here (this is where sst_rest is used):

veros/veros/setups/north_atlantic/north_atlantic.py

Lines 329 to 335 in c638ac0

	vs.forc_temp_surface = (
	(f1 * vs.sst_rest[:, :, n1] + f2 * vs.sst_rest[:, :, n2])
	* (f1 * vs.sst_clim[:, :, n1] + f2 * vs.sst_clim[:, :, n2] - vs.temp[:, :, -1, vs.tau])
	* vs.maskT[:, :, -1]
	/ cp_0
	/ settings.rho_0
	)

That parameter is also often called sea surface temperature relaxation time scale. I can't really help you with choosing this value (not an expert in that field). I apologize for the poor user experience; spinning up custom setups from scratch is not something we do a lot, so thanks for being patient.